Produced by the Science/AAAS Custom Publishing Ofce

Autonomic Neuroscience Centre, University College Medical School, London, UK,and Department of Pharmacology and Therapeutics, The University of Melbourne,Melbourne, Australia

g.burnstock@ucl.ac.uk

Purinergic signaling in acupuncture

The proposed role of puriner-gic signaling in the physiological

basis of acupuncture was Brst presented in 2009.Data showing that ATP is released from kerati-nocytes and other skin cells during acupuncturetreatments lends weigh to this hypothesis. ATPin turn activates P2X3 receptors on the sensorynerves in the skin, which then transmit thosemessages to motor neurons in the brain stemthat control autonomic functions and modulatenociceptive activities. Here, we review and de-scribe the recent evidence for purinergic signalingunderlying acupuncture effects and propose waysto further test this hypothesis.

Introduction

It has been well established that adenosine5’-triphosphate (ATP) is an intracellular energysource in cellular biochemistry. In 1970, Burnstocket al. suggested that ATP acted as a nonadrener-gic, noncholinergic neurotransmitter in the gut (1),and in 1972 he named the extracellular actions ofATP, “purinergic signaling” (since ATP is a purinenucleotide), and formulated the purinergic signal-ing hypothesis (2).

In 2009, Burnstock proposed that purinergicsignaling could be involved in the physiologicalmechanisms mediating acupuncture effects.This hypothesis suggested that mechanicaldeformation of the skin by needles or applicationof heat or electrical current leads to the releaseof large amounts of ATP from keratinocytes,Bbroblasts, and other cell types in skin (Figure 1).The released ATP then activates P2X3 ion channelreceptors on sensory nerves within the skinand tongue that transmit messages via sensoryganglia and the spinal cord to the brain stemand hypothalamus. These brain regions containmotor neurons that control autonomic functions,including cardiovascular, gastrointestinal,respiratory, and urinogenital activities—commontargets of acupuncture treatments. These sensoryneuron messages also modulate the pathwaysthat lead to centers in the cortex responsiblefor conscious awareness of pain and othercentral nervous system activities, including sleepregulation (3). A number of subsequent studies have beenpublished that also implicate purinergic signaling in variousaspects of acupuncture, detailed below.

Materials that appear in this section were not reviewed orassessed by Science Editorial staff, but have been evaluated byan international editorial team consisting of experts in traditionalmedicine research.

Author:

Geoffrey Burnstock

Supporting evidence for the hypothesisStudies that have established the components involved

in the purinergic signaling pathway include: (i) release ofATP (in response to mechanical or chemical stimulation)

FIGURE 1. Acupuncture

and purinergic

signaling. Insertion

and twisting of the

needles employed

in acupuncture

mechanically deforms

the skin, leading to

the release of ATP by

skin keratinocytes (1).

ATP binds to specifc

receptors located on

sensory nerve endings

in the skin, P2X3

and P2X2/3 (2). The

signaling message is

then relayed via dorsal

root ganglia to the

spinal cord (3) and

subsequently through

interneuronal pathways

(4) to the brain stem

(5) that contains motor

neurons, which control

the functions of gut,

lung, heart, arteries,

and reproductive

organs—all major

targets for acupuncture.

Signals also travel to

the pain centers of

the cortex, delivering

a message to inhibit

pain (6). (Reproduced

with permission from

reference 36.)

S23

Produced by the Science/AAAS Custom Publishing Ofce

S24

from keratinocytes (4–6) and possibly from Merkel cells,which contain high levels of ATP (7, 8); ATP has also beenshown to be released from keratinocytes upon heating (9);(ii) immunohistochemical data demonstrating the presenceof P2X3 receptors on sensory nerve fbers in the skin (10–12)and tongue (13); (iii) in an isolated tongue/lingual nervepreparation, mechanical activation of the tongue with DeFrey hairs was shown to result in a discharge in the lingualsensory nerve fbers that was mimicked by ATP activationand blocked by P2X3 receptor antagonists (14); and (iv) bothpresynaptic inhibition via adenosine A

1and P2Y receptors,

and enhancement via P2X and A2Areceptors at synapses in

the central nervous system have been reported (15).Subsequent papers have built upon and extended

evidence in support of purinergic signaling underlyingacupuncture effects. Several studies have associated the skincells affected by acupuncture techniques with purinergicsignaling. For example, ATP has been shown to be releasedfrom human keratinocytes in response to mechanicalstimulation by hypo-osmotic shock (16), as well as fromkeratinocytes in response to heat (17). Additionally, mastcells, which accumulate around the acupuncture needles,also release ATP in response to mechanical stimulation (18).Another skin cell type, human subcutaneous fbroblasts, can

FIGURE 2. Schematic illus-

tration of the central neural

pathways that carry afferent

(sensory) neural impulses

following acupuncture

treatment from various parts

of the body. Brain areas

that commonly respond in

neuroimaging studies to

acupuncture stimulation are

indicated with gray shadow.

DCEAS: dense cranial elec-

troacupuncture stimulation.

(Reproduced from 28, with

permission from Hindawi

Publishing Corporation.)

Produced by the Science/AAAS Custom Publishing Ofce

S25

also release ATP in response to bradykinin and histamine (19,20). Tsutsumi et al. demonstrated that mechanical stimulationcan evoke the propagation of calcium waves betweenhuman keratinocytes, induced by ATP and activation ofP2Y

2receptors (21, 22), which is consistent with the earlier

results from Koizumi et al. (5). Tuina (traditional therapeuticmassage) and moxibustion (a traditional Chinese medicinetherapy using a moxa, often made from dried mugwort,either used as a Muff or processed into a cigar-shaped stick; itcan be used indirectly, with acupuncture needles, or burnedon to the patient’s skin) may also act via the purinergicsignaling pathway (23). Papers describing the release ofATP from human epidermal keratinocytes via connexinhemichannels and vesicles involving vesicular nucleotidetransporter have recently been published (24–26). A 2010study has claimed that adenosine, following breakdown ofreleased ATP during acupuncture, can act as a prejunctionalinhibitor of neurotransmission via A

1receptors, resulting in

anti-nociceptive actions (27). Valuable reviews are availabledescribing the neural pathways from different skin regions tostructures in the brain stem and higher brain centers. Thesepathways are important because different acupuncture sitesmay activate different neural pathways impinging on specifcnuclei in the brain stem that control autonomic functionspotentially modulated by acupuncture (Figure 2) (28, 29).

Purinergic signaling and electroacupunctureElectroacupuncture is a form of acupuncture where a

small electric current is passed between pairs of acupunctureneedles. This is thought to augment traditional acupunctureand is believed to be particularly helpful in treating pain.

The supraspinal antinociception effect ofelectroacupuncture has been associated with P2X3 receptoractivation in the midbrain periaqueductal gray region (30).Moreover, the analgesic effect of electroacupuncture onchronic neuropathic pain has been shown to be mediatedby P2X3 receptors in rat dorsal root ganglion neurons (31).Following these studies, electroacupuncture was shown toresult in a reduced expression of P2X3 and P2X2 receptors inthe dorsal root ganglion of rats with chronic neuropathic pain(32) and visceral hypersensitivity (33). Electroacupunctureat He-Mu points can also reduce P2X4 receptor expressionin colon and spinal cord in visceral hypersensitivity (34).Moreover, in a review by Lin et al., the neuroprotectiveeffects of acupuncture were reported to act via increasingbrain derived neurotrophic factor (BDNF) expression viastimulation of ATP (35).

ConclusionsEvidence in support of the hypothesis of purinergic

signaling mediating the physiological mechanismsunderlying acupuncture effects has been accumulating overrecent years. To help further test this hypothesis, I proposethat experienced acupuncturists focus on acupuncturesites that induce effects that can be quantifed, such as an

increase or decrease in heart rate or blood pressure, andidentify specifc neurons that are activated in the brain usingnoninvasive scanning techniques. If acupuncture-inducedeffects can be identifed and quantifed, researchers couldthen test whether ATP mimicks the responses and if P2X3receptor antagonists block the effects. Moreover, wesuggest that researchers conduct experiments recordingresponses from sensory neurons in the skin and tonguein animal models and distinguish between low-thresholdfbers involved in acupuncture and high-threshold fbersthat mediate nociception, as well as recordings from themotor nerves in the brainstem responsible for autonomicfunctions.

References

1. G. Burnstock, G. Campbell, D. Satchell, A. Smythe, Br. J. Pharmacol. 40, 668 (1970).

2. G. Burnstock, Pharmacol. Rev. 24, 509 (1972).3. G. Burnstock,Med. Hypotheses 73, 470 (2009).4. N. Mizumoto, M. E. Mummert, D. Shalhevet, A. Takashima, J.

Invest. Dermatol. 121, 1066 (2003).5. S. Koizumi et al., Biochem. J. 380, 329 (2004).6. H. E. Burrell et al., J. Biol. Chem. 280, 29667 (2005).7. R. Crowe, M.Whitear, Cell Tissue Res. 190, 273 (1978).8. M. Silberstein,Med. Hypotheses 75, 272 (2010).9. S. Mandadi et al., Pfugers Arch. 458, 1093 (2009).10. E. J. Bradbury, G. Burnstock, S. B. McMahon,Mol. Cell. Neurosci.

12, 256 (1998).11. G. Burnstock, Br. J. Anaesth. 84, 476 (2000).12. M. Taylor, J. C. Peleshok, A. Ribeiro-da-Silva, J. Comp. Neurol.

514, 555 (2009).13. X. Bo et al.,Neuroreport 10, 1107 (1999).14. W. Rong, G. Burnstock, K. M. Spyer, J. Physiol. 524, 891 (2000).15. G. Burnstock, Physiol. Rev. 87, 659 (2007).16. N. Azorin et al., Exp. Dermatol. 20, 401 (2011).17. J. R. Gifford, C. Heal, J. Bridges, S. Goldthorpe, G.W. Mack, J.

Physiol. 590, 6403 (2012).18. L. Wang et al., Evid. Based Complement. Alternat. Med. 2013,

350949 (2013).19. R. Pinheiro et al., Cell Commun. Signal. 11, 70 (2013).20. R. Pinheiro et al., J. Biol. Chem. 288, 27571 (2013).21. M. Tsutsumi et al., Cell Tissue Res. 338, 99 (2009).22. M. Tsutsumi et al., Skin Res. Technol. 16, 146 (2010).23. L. Fan, L. M. Yin, J. Acupunct. Tuina Sci. 12, 125 (2014).24. T. P. Barr et al., PLOS ONE 8, e56744 (2013).25. T. Takahashi et al., J. Invest. Dermatol. 133, 2407 (2013).26. K. Inoue et al., J. Invest. Dermatol. 134, 1465 (2014).27. N. Goldman et al.,Nat. Neurosci. 13, 883 (2010).28. Z. J. Zhang, X. M.Wang, G. M. McAlonan, Evid. Based

Complement. Alternat. Med. 2012, 429412 (2012).29. Q. Q. Li et al., Evid. Based Complement. Alternat. Med. 2013,

267959 (2013).30. Z. Xiao et al., Brain Res. 1330, 31 (2010).31. W. Z. Tu et al.,Neurochem. Int. 60, 379 (2012).32. R. D. Cheng et al., Chin. J. Integr. Med. 19, 374 (2013).33. Z. Wang et al.,Neural Regen. Res. 8, 802 (2013).34. X. Guo et al.,Neural Regen. Res. 8, 2069 (2013).35. D. Lin et al., Int. J. Mol. Sci. 15, 3234 (2014).36. G. Burnstock, The Scientist 25, 24 (2011).